Prazto Inc. (Headquarters: Chuo-ku, Tokyo; Representative Director: Satoshi Haga), a provider of the data integration platform "Passwork," will launch "Passwork MCP," an AI integration function that allows users to build, execute, and operate business data integrations from conversations on Slack, on July 1, 2026 (Wed.). It converts natural conversation instructions exchanged on Slack into verified and reproducible data integration processes (deterministic data integration) that can be executed as production operations. This achieves both the ease of conversation and the accuracy, reproducibility, and auditability required for production operations. ◼️ Background: True Business Efficiency Lies in "Daily, Granular Tasks" While large-scale automation often garners attention for business efficiency, what truly consumes the time of on-site staff are the small data tasks repeated daily – tasks like master data reconciliation, data preparation before reporting, and data entry between systems, all of which are a cumulative effort of seemingly minor operations. However, it has been difficult to extend automation to these granular tasks. Traditional automation is cumbersome to build, making it impractical to apply to each small task individually. On the other hand, while generative AI offers ease of use, its outputs can be inconsistent or unpredictable, making it unsuitable for core data or production operations. These two barriers have confined efficiency improvements to only some large-scale processes. True efficiency is only achieved when reliable integration permeates each of these "daily, granular tasks." To accomplish this, a mechanism was needed that was as easy as conversation but possessed the accuracy required for production. ◼️ Three Features of "Passwork MCP" 【Point 1】Just Converse. Execution is "Verified Deterministic Data Integration" AI does not "judge the process on the spot." The AI's role is limited to understanding intent, and Passwork converts that intent